For those still using the 11th gen FL13 with their original (first) battery since 2021/2022, if it’s been more or less your daily driver (700+ charges by now), what’s your battery’s current capacity?
4 Likes
EDIT: Turns out /sys/class/power_supply/BAT1 is a dirty liar. Original post preserved below:
My battery is from the 11th gen batch 5 received in early November 2021. It has been my daily driver, but it spent most of its life plugged in, so my battery is reporting far fewer cycles, and I enabled the battery charge limiter as soon as that was added to the 11th gen firmware.
(for this output I set my limit back to 100%, charged it up to full, unplugged it, and let it drain a little because idk how it determines
charge_full. No idea if that accomplished anything.)$ find /sys/class/power_supply/BAT1/ -maxdepth 1 -type f -name 'capacity*' -or -name 'charge*' -or -name 'cycle*' -or -name 'manufacturer' -or -name 'model_name' -or -name 'voltage_min_design' | while read f; do echo "/sys/class/power_supply/BAT1/$(basename "$f")" $(cat "$f"); done | sort /sys/class/power_supply/BAT1/capacity 97 /sys/class/power_supply/BAT1/capacity_level Normal /sys/class/power_supply/BAT1/charge_control_end_threshold 100 /sys/class/power_supply/BAT1/charge_full 2923000 /sys/class/power_supply/BAT1/charge_full_design 3572000 /sys/class/power_supply/BAT1/charge_now 2844000 /sys/class/power_supply/BAT1/cycle_count 222 /sys/class/power_supply/BAT1/manufacturer NVT /sys/class/power_supply/BAT1/model_name Framewo /sys/class/power_supply/BAT1/voltage_min_design 15400000So it looks like my battery still has 81.83% of its original capacity based on
2923000/3572000 * 100
3 Likes
Nice. What % was your charge limit set to? And does your CPU boost / under load often (usage pattern)?
It was set to 70%.
Back when I had the Intel motherboard, I disabled turbo boost through /sys/devices/system/cpu/intel_pstate/no_turbo and I set the hwpstates to favor energy efficiency over performance through /sys/devices/system/cpu/cpufreq/policy*/energy_performance_preference.
Now that I am on the AMD 7840U board I don’t have turbo disabled (because I haven’t figured out how) but I still set the same power setting for the energy_performance_preference and I set /sys/firmware/acpi/platform_profile to low-power.
Usage is mostly coding, compiling that code, and occasionally factorio. It spends more time plugged into a thunderbolt 4 dock than it spends unplugged.
1 Like
Really appreciate the detailed response. With that info, I can start using mine as a daily driver then. The battery longevity has been a concern of mine due to to micro charge / discharge cycle behaviour. But with your long term, mostly plugged in, no turbo usage pattern, that now gives me higher degree of assurance that the battery can last me 4+ years.
2 Likes
find /sys/class/power_supply/BAT1/ -maxdepth 1 -type f -name 'capacity*' -or -name 'charge*' -or -name 'cycle*' -or -name 'manufacturer' -or -name 'model_name' -or -name 'voltage_min_design' | while read f; do echo "/sys/class/power_supply/BAT1/$(basename "$f")" $(cat "$f"); done | sort
/sys/class/power_supply/BAT1/capacity 100
/sys/class/power_supply/BAT1/capacity_level Normal
/sys/class/power_supply/BAT1/charge_full 3041000
/sys/class/power_supply/BAT1/charge_full_design 3572000
/sys/class/power_supply/BAT1/charge_now 3041000
/sys/class/power_supply/BAT1/cycle_count 142
/sys/class/power_supply/BAT1/manufacturer NVT
/sys/class/power_supply/BAT1/model_name Framewo
/sys/class/power_supply/BAT1/voltage_min_design 15400000
My Power Manager gives me 85% of real capacity.
I don’t remember the batch number, but I bought my FW13 in May2022. Laptop is mostly plugged in and I set the charge limit for some time at 90% (but since I had to replace the CMOS battery, I didn’t set it back). However I didn’t disable the turbo.
You’ll probably get more accurate data by fully discharging it and then charging it to 100.
4 Likes
Ah yeah, I didn’t bother discharging it first because I let it discharge fully ~2 days ago (unintentionally). But I had my charge limiter still enabled so I charged my laptop to 70% where it stayed for the past 2 days before I let it go up to 100 last night.
Sounds like it has seen min and max relatively recently so the data is probably pretty solid.
1 Like
It’s recommended to use ectool since on other programs the cycle count is reset each 256. Recently I borrowed an earlier battery that has 1134 cycles on ectool battery but both tlp-stat -b and upower -i $(upower -e | grep battery) show 64
1 Like
Have you done a full charge / discharge / charge cycle recently?
A couple of weeks ago I think but I can do one today.
I was monitoring the capacity because I was thinking of replacing the battery by a 61WH if it ever went too low but it looks like I still have time before doing so 
.
Whoa, you’re right. Turns out /sys/class/power_supply/BAT1 is a dirty liar. Here’s my output from ectool which shows over 3x the cycles that /sys/class/power_supply/BAT1 shows:
$ sudo ectool battery
Battery 0 info:
OEM name: NVT
Model number: Framework L
Chemistry : LION
Serial number: 014E
Design capacity: 3572 mAh
Last full charge: 2906 mAh
Design output voltage 15400 mV
Cycle count 735
Present voltage 16101 mV
Present current 0 mA
Remaining capacity 2036 mAh
Desired voltage 17400 mV
Desired current 3572 mA
Flags 0x0b AC_PRESENT BATT_PRESENT CHARGING
The remaining capacity seems to have been approximately correct though, since this is showing 81.35% from 2906/3572 * 100.
1 Like
I had to replace it last month with the new battery, since I didn’t set a charging limit on it and it was noticeably not lasting as long (I didn’t take measurements, sorry)
So I can definitely confirm you do not want to charge it to 100%
I’m at about 86%, it seems (my Framework is from late 2021 as well, mostly plugged in to my hub):
$ find /sys/class/power_supply/BAT1/ -maxdepth 1 -type f -name 'capacity*' -or -name 'charge*' -or -name 'cycle*' -or -name 'manufacturer' -or -name 'model_name' -or -name 'voltage_min_design' | while read f; do echo "/sys/class/power_supply/BAT1/$(basename "$f")" $(cat "$f"); done | sort
/sys/class/power_supply/BAT1/capacity 100
/sys/class/power_supply/BAT1/capacity_level Normal
/sys/class/power_supply/BAT1/charge_control_end_threshold 100
/sys/class/power_supply/BAT1/charge_full 3087000
/sys/class/power_supply/BAT1/charge_full_design 3572000
/sys/class/power_supply/BAT1/charge_now 3087000
/sys/class/power_supply/BAT1/cycle_count 11
/sys/class/power_supply/BAT1/manufacturer NVT
/sys/class/power_supply/BAT1/model_name Framewo
/sys/class/power_supply/BAT1/voltage_min_design 15400000
$ sudo ectool battery
EC result 3 (INVALID_PARAM)
Battery info:
OEM name: NVT
Model number: Framewo
Chemistry : LION
Serial number: 0037
Design capacity: 3572 mAh
Last full charge: 3087 mAh
Design output voltage 15400 mV
Cycle count 267
Present voltage 17239 mV
Present current 0 mA (discharging)
Remaining capacity 3088 mAh
Flags 0x07 AC_PRESENT BATT_PRESENT DISCHARGING
I’m still on the original battery from a Framework 11th gen, batch 6 at the end of 2021. ectool says I’m at 552 cycles. I used the laptop pretty lightly the first year or something, and I’ve used it fairly heavily since then, on and off the charger. Also went through setting a variety of different max charge levels, between 80% and 100%.
My reported capacity varies, and I haven’t cycled the battery lately, but it seems like I’m between 82% and 84% someplace.
Battery info:
OEM name: NVT
Model number: Framewo
Chemistry : LION
Serial number: 0190
Design capacity: 3572 mAh
Last full charge: 3098 mAh
Design output voltage 15400 mV
Cycle count 552
Present voltage 17073 mV
Present current 0 mA
Remaining capacity 2914 mAh
Flags 0x0b AC_PRESENT BATT_PRESENT CHARGING
@Tom1 Where’d you get your ectool? Mine doesn’t show a Desired Voltage.
Looks like mine is from Dustin L. Howett / ectool · GitLab which I built from the AUR back in Oct 2024. No idea how it compares to other copies/versions of ectool.
$ doas pacman -Qi ectool
Name : fw-ectool-git
Version : r2763.0ac6155-1
Description : ectool for the Framework laptop.
Architecture : x86_64
URL : https://gitlab.howett.net/DHowett/ectool
Licenses : BSD-3-Clause
Groups : None
Provides : ectool
Depends On : libftdi
Optional Deps : None
Required By : None
Optional For : None
Conflicts With : None
Replaces : None
Installed Size : 247.80 KiB
Packager : Unknown Packager
Build Date : Wed Oct 9 15:10:34 2024
Install Date : Wed Oct 9 15:11:00 2024
Install Reason : Explicitly installed
Install Script : No
Validated By : SHA-256 Sum Signature
1 Like
$ git remote -v
origin https://github.com/FrameworkComputer/EmbeddedController (fetch)
origin https://github.com/FrameworkComputer/EmbeddedController (push)
… huh.
Every time I try to dive into git forks and branches I wind up getting entangulated, too-weak git skills and/or missing context from not working on the projects.
It’s a bit weird to me that a Chromium OS-derived repo is ahead of and more featureful than the Framework ectool, on a Framework laptop. I dunno, I guess it builds, seems to run… cool.